Chunkr API Client

A functional TypeScript client for the Chunkr.ai document processing API.

Features

• Create document processing tasks from URLs or base64-encoded files
• Poll task status until completion
• Retrieve and filter tasks
• Extract structured document chunks and metadata
• Complete error handling and retries
• Task management (delete, cancel, update)
• API health monitoring

Installation

npm install @legalflow/chunkr-client

Configuration

Add your Chunkr API key to the .env file:

CHUNKR_API_KEY=sk-xxxxxxxxxxxx

The API URL is set internally to https://api.chunkr.ai/api/v1 by default.

Usage Examples

Basic Document Processing Flow

import { runChunkrDocumentFlow } from '@legalflow/chunkr-client';

// Process a document from a URL
const processDocument = async (documentUrl: string, fileName: string) => {
  try {
    const result = await runChunkrDocumentFlow(documentUrl, fileName);

    console.log(`Processed document with ${result.chunks.length} chunks`);
    console.log(`Document metadata:`, result.metadata);

    // Use the chunks for further processing
    return result.chunks;
  } catch (error) {
    console.error('Document processing failed:', error);
    throw error;
  }
};

// Example usage:
processDocument('https://example.com/document.pdf', 'important-contract.pdf')
  .then(chunks => {
    // Do something with the chunks
  })
  .catch(error => {
    // Handle errors
  });

Processing Base64-Encoded Documents

import { runChunkrDocumentFlowFromBase64 } from '@legalflow/chunkr-client';

const processBase64Document = async (
  base64Content: string,
  fileName: string
) => {
  try {
    const result = await runChunkrDocumentFlowFromBase64(
      base64Content,
      fileName
    );
    return result.chunks;
  } catch (error) {
    console.error('Base64 document processing failed:', error);
    throw error;
  }
};

Custom API URL

import { createChunkrApiClient } from '@legalflow/chunkr-client';

// Create a client with a custom API URL
const customClient = createChunkrApiClient('https://custom-chunkr-api.example.com/v1');

Advanced Configuration

import { runChunkrDocumentFlow } from '@legalflow/chunkr-client';

// Using advanced configuration options
const result = await runChunkrDocumentFlow(documentUrl, fileName, {
  ocr_strategy: 'All',
  high_resolution: true,
  chunk_processing: {
    ignore_headers_and_footers: true,
    target_length: 1024 // Longer chunks
  }
});

API Documentation

For detailed API documentation, please refer to the official Chunkr API Documentation.

License

MIT